Exploiting Korean Language Model to Improve Korean Voice Phishing Detection

초록

Text classification task from Natural Language Processing (NLP) combined with state-of-the-art (SOTA) Machine Learning (ML) andDeep Learning (DL) algorithms as the core engine is widely used to detect and classify voice phishing call transcripts. While numerousstudies on the classification of voice phishing call transcripts are being conducted and demonstrated good performances, with the increaseof non-face-to-face financial transactions, there is still the need for improvement using the latest NLP technologies. This paper conductsa benchmarking of Korean voice phishing detection performances of the pre-trained Korean language model KoBERT, against multipleother SOTA algorithms based on the classification of related transcripts from the labeled Korean voice phishing dataset called KorCCVi. The results of the experiments reveal that the classification accuracy on a test set of the KoBERT model outperforms the performancesof all other models with an accuracy score of 99.60%.

키워드